We recently described a new design for a molecular-based memory device wherein a layer of redox-active molecules is tethered to an ultra-thin dielectric surface, which in turn is deposited on a semiconductor. 1 The dielectric layer and the molecular tether (linker and surface attachment group) both provide barriers to electron transfer between the semiconductor and the redox-active molecule. In this type of molecular-based field effect transistor, the charge stored in the molecules can change the current level in the transistor, thereby affording a non-destructive means by which the charge state of the molecules can be detected.
In contrast with conventional semiconductor-based devices, the use of charge-storage molecules exploits the power of synthetic design to tailor molecules that operate at low voltage and that provide multiple charged states. Great latitude also exists in the design of the barriers presented by both the tether and the dielectric layer. The barrier presented by the tether can be tuned via synthetic organic chemistry while that of the dielectric can be tuned by semiconductor-processing techniques. In particular, the composition (and length) of the tether can be varied from insulating aliphatic groups to more conducting conjugating species. Likewise, composition of the dielectric layer can be a commonly used SiO2 layer or a metal oxide such as HfO2, ZrO2, etc.
Our preliminary studies employed ferrocenylmethylphosphonic acid as the charge-storage molecule.1 The phosphonic acid group anchors the charge-storage molecule to the oxide surface. The initial success of this approach has prompted us to investigate the synthesis of a much wider variety of charge-storage molecules, particularly porphyrinic molecules, which bear phosphonic acid-terminated linkers. Porphyrins bearing phosphonic acid tethers have been synthesized and attached to oxide surfaces for a variety of other applications including solar energy, oxidative catalysis, sensing, and recognition of polysaccharides.2-14 
The synthetic approaches that have been employed to prepare porphyrins bearing phosphonic acid/phosphonate units can be characterized by (1) whether the phosphonate unit is introduced into precursors to the porphyrin or by derivatization of a preexisting porphyrin, (2) whether statistical or rational routes are employed, (3) the number and pattern of phosphonate groups at the perimeter of the porphyrin, (4) the type of phosphonic acid protecting group employed, (5) the nature of the central metal, and (6) the method of cleavage of the phosphonate protecting groups.
A4-Porphyrins bearing four arylphosphonic acids have been prepared by condensation of a dialkoxyphosphorylbenzaldehyde with pyrrole followed by deprotection of the free base porphyrin.2 Alternatively, the free base porphyrin can be metalated followed by deprotection.4,5 A4-porphyrins bearing four alkylphosphonic acids have been prepared by derivatization of a reactive halo-substituted porphyrin.5,7 A3B-porphyrins bearing a single phosphonic acid have been prepared by a mixed-aldehyde condensation of a dialkoxyphosphorylbenzaldehyde, benzaldehyde, and pyrrole;4 or by derivatization of a porphyrin bearing a single reactive halo group.6,14 Trans-A2B2-porphyrins bearing two phosphonic acid groups have been prepared by condensation of a dialkoxyphosphorylbenzaldehyde and dipyrromethane.12 Chlorins bearing two phosphonic acids have been prepared by derivatization of a deuterochlorin-dibromide with tris(trimethylsilyl)phosphite.9 In each case, the porphyrinic species were employed as the free base or as a metal chelate that is rather robust toward the acidic conditions for cleavage of the dialkyl phosphonate. The metals include Mn,4,5 Fe,9 Co,9 Ni,9 Pd,6 and Os,7 which are all categorized in the porphyrin field as class I or class II metals, affording chelates that are exceptionally resilient toward acids.15 In general, phosphonic acids combine with metals to give extended, often insoluble, metal phosphonates. A rare case wherein metalation was performed in the presence of a free phosphonic acid employed a porphyrin superstructure containing a hindered phosphonic acid.14 
One of the considerable attractions of molecular information storage is the ability to tune the properties of the charge-storage molecules through molecular design. In studies of thiol-derivatized porphyrins, we found that the period during which the oxidized molecules remained charged (i.e., the charge-retention time) depends quite sensitively on the length of the tether (linker and surface attachment group). For example, as the number of methylene groups in the tether phenyl-(CH2)n—S— increased along the series 0, 1, 2, and 3, the charge-retention time increased from 116, 167, 656 to 885 s. The rate of electron-transfer (reading process) also slowed with increase of linker length. Moreover, the quality (uniformity, integrity) of the self-assembled monolayers (SAMs) increased in going from the phenylthio tether (n=0) to the phenylalkylthio tethers (n=1-3).16 